I suppose an online repository (@xen-audio.com) featuring the latest version of every relevant document makes more sense.
Don't think you'll avoid the "which version/where can I find it/etc" questions though!
But at least you'll be able to copy-paste the answer, have it as a forum signature or something.
Don't think you'll avoid the "which version/where can I find it/etc" questions though!
But at least you'll be able to copy-paste the answer, have it as a forum signature or something.
Another PCB for ES9023 + JG Filter Buffer
Dear all,
there are already very nice versions of ES9023 and JG filter buffer PCBs, however I wanted to combine both and do my own (its called DIYaudio, isn't it?). Also the combination of I2S input and JG filter buffer is currently not available (Acko has one but JG Buffer is stacked).
My design is definitely inspired by existing schematics/boards, but also by playing around with different applications of ES9023 using a SOP/DIP adapter (it's really easy to get already quite good results that way so I can only second EUVLs recommendation: just do it, it's fun).
I am about to finalize the board (some fine tuning here and there) and will order protos soon. Any comments are welcome, especially regarding the routing of ground. I used a solid groundplane without any interrupt, however it's not separated into analog ground and digital ground and currently I don't have ground planes on top layer.
Best regards, Daniel
Dear all,
there are already very nice versions of ES9023 and JG filter buffer PCBs, however I wanted to combine both and do my own (its called DIYaudio, isn't it?
). Also the combination of I2S input and JG filter buffer is currently not available (Acko has one but JG Buffer is stacked).My design is definitely inspired by existing schematics/boards, but also by playing around with different applications of ES9023 using a SOP/DIP adapter (it's really easy to get already quite good results that way so I can only second EUVLs recommendation: just do it, it's fun
).I am about to finalize the board (some fine tuning here and there) and will order protos soon. Any comments are welcome, especially regarding the routing of ground. I used a solid groundplane without any interrupt, however it's not separated into analog ground and digital ground and currently I don't have ground planes on top layer.
Best regards, Daniel
Attachments
Thanks
When I order some boards I would be pleased to make some available for "beta test" (no guarantee of beeing free from any error). I am also generally open for a later group buy if there is interest, however first I would want to talk to Joachim Gerhard and Jean Paul/Subbu since I don't want to get in competition in any way. Should be just an alternative for those who want to have (only) I2S in + JG filter buffer.
I used Panasonic ECHU PPS Film caps for C3 and C4 with very good results (same for C43 and C53 as per the JG filter buffer group buy BOM). However maybe FKP will additionally fit as an option.
ADP-150 really have impressive specs noise wise (more than 6dB less than LT1761) but they are not available as 3.6V version. Feeding the ES9023 with 3.6V gave slightly better results in my tests and others reported the same. But main reason was easy sourcing of LT1761 which is available at Reichelt Wanted to source all parts at a maximum of two suppliers (Mouser, Reichelt) to save shipping cost. Complete BOM isn't that much. Maybe another supplier could deliver everything?
Best regards, Daniel
When I order some boards I would be pleased to make some available for "beta test" (no guarantee of beeing free from any error
). I am also generally open for a later group buy if there is interest, however first I would want to talk to Joachim Gerhard and Jean Paul/Subbu since I don't want to get in competition in any way. Should be just an alternative for those who want to have (only) I2S in + JG filter buffer.I used Panasonic ECHU PPS Film caps for C3 and C4 with very good results (same for C43 and C53 as per the JG filter buffer group buy BOM). However maybe FKP will additionally fit as an option.
ADP-150 really have impressive specs noise wise (more than 6dB less than LT1761) but they are not available as 3.6V version. Feeding the ES9023 with 3.6V gave slightly better results in my tests and others reported the same. But main reason was easy sourcing of LT1761 which is available at Reichelt
Wanted to source all parts at a maximum of two suppliers (Mouser, Reichelt) to save shipping cost. Complete BOM isn't that much. Maybe another supplier could deliver everything? Best regards, Daniel
Add 4 pads & holes then
I would also use an indipendent reg for the flip-flops
(and mod some tracks to be at 45°
)Digikey is selling Analog Device as long as I know, but I understand your point of view (they charge 18€ for small orders) They also sell susumu resistors and pana pps caps cheaper than mouser
Last edited:
adds another unintended inductor, that goes for the i2s inputs too.
nice job curryman, i'll have a bit more of a look, but the only thing I see at the moment is the clock and i2s trace routing and impedance. the clock traces seem too thin to me, for a 2 layer board they need to be thicker than this for 50ohms and thicker is lower noise for fast signals. the exact thickness will depend on the thickness of the PCB substrate.
also wrt the u.fl connectors, again having the pinheader after the u.fl like that, adds either an inductor if the header isnt fitted as well as an aerial if it is.
try doing it like this. the trace thickness in these pics isnt calculated, it was a rough illustration I did for someone else doing the same thing.
Attachments
Last edited:
Thank you all for your valuable comments
Please find attached the latest version of the board. I added another reg for the flipflops (By the way: soldering those tiny parts will be quite tough). Since LT1761 and ADP-150 are pin compatible both can be used. Just add a C at the input for ADP or the bypass C for LT1761. Kept the LT1761-Adj for ES9023 power supply to allow for 3.6V though .
I removed the series resistors at the I2S line receivers but added some at source sides (did some reading and Popolvars comment makes sense). I also changed pin headers and U.FL connectors as per qusp's recommendation. I2S traces are 0.7mm.
Did not change the output filter caps (C3 and C4). The Wimas are quite big and I didn't feel comfortable with the resulting layout. Hope you don't mind (the PPS caps are really good ).
Any further comments are appreciated!
Please find attached the latest version of the board. I added another reg for the flipflops (By the way: soldering those tiny parts will be quite tough
). Since LT1761 and ADP-150 are pin compatible both can be used. Just add a C at the input for ADP or the bypass C for LT1761. Kept the LT1761-Adj for ES9023 power supply to allow for 3.6V though .I removed the series resistors at the I2S line receivers but added some at source sides (did some reading and Popolvars comment makes sense
). I also changed pin headers and U.FL connectors as per qusp's recommendation. I2S traces are 0.7mm.Did not change the output filter caps (C3 and C4). The Wimas are quite big and I didn't feel comfortable with the resulting layout. Hope you don't mind (the PPS caps are really good
).Any further comments are appreciated!
Attachments
Last edited:
I wasnt suggesting resistors (notice I didnt mention them) but that doesnt mean you shouldnt have them, its a build by build thing and some situations will indeed call for termination at both ends.
like I said, the pics were made to illustrate for another thread, the resistor network was just a neat way to tie the traces to something, as you cant have traces leading into thin air in diptrace. the pics were to illustrate the u.fl + header thing only
like I said, the pics were made to illustrate for another thread, the resistor network was just a neat way to tie the traces to something, as you cant have traces leading into thin air in diptrace. the pics were to illustrate the u.fl + header thing only
Aren't that big if you move Out_L and Out_R aside (or remove as you most probably won't use them as you integrated the JG Filter)
I can still see some tracks not at 45° or 90°.... just for aesthetic
qusp, how to measure and design the source and receiver resistors? My old 10 MHz Hameg might help for LRCLK but for MCLK (e.g. 49.152 MHz) even a 100 MHz scope only shows 1st and 2nd harmonic? I guess not many will be able to test and design suitable values for their specific application
And what about the effect of having 0R resistors compared to just a trace?
If I want to have 50 Ohm impedance for the clock lines, TXline gives me 0.75mm trace width (1mm FR4, 35µm, 0.15mm gap) if I have a solid ground plane on bottom layer and surround the lines with GND on top layer. Would you agree?
Just as a small upgrade another version
And what about the effect of having 0R resistors compared to just a trace?
If I want to have 50 Ohm impedance for the clock lines, TXline gives me 0.75mm trace width (1mm FR4, 35µm, 0.15mm gap) if I have a solid ground plane on bottom layer and surround the lines with GND on top layer. Would you agree?
Just as a small upgrade another version
Attachments
Thanks
When I order some boards I would be pleased to make some available for "beta test" (no guarantee of beeing free from any error
I used Panasonic ECHU PPS Film caps for C3 and C4 with very good results (same for C43 and C53 as per the JG filter buffer group buy BOM). However maybe FKP will additionally fit as an option.
ADP-150 really have impressive specs noise wise (more than 6dB less than LT1761) but they are not available as 3.6V version. Feeding the ES9023 with 3.6V gave slightly better results in my tests and others reported the same. But main reason was easy sourcing of LT1761 which is available at Reichelt
Best regards, Daniel
what about using LT1763
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.